RX 580 8gb unstable with factory settings

[Bowkiller84]

Hello fellow tech juckies. I have recently rebuilt my rig (3rd gen in 12 years). And I’m having some trouble with my GPU crashing. Hoping for some help. I have tried a handful of ideals already but any help would be appreciated. 

 

So to start here is my rig specs 

 

AMD Ryzen 5 1600 with a stable OC at 3.95ghz (has stress test with aida64 for over 9hours on 3 different Occasions) cooled by a deepcool captain 240 with max temps of 72c and normal load temps in the low 50c

 

g-skill trident Z ddr4 ram 3600mhz OC to 3333 stable as well with the same stress test and times as the CPU

 

MSI B 450 tomahawk motherboard

 

Corsair rm850i PSU (under full load using just under 400w give or take) 

 

kington SSD boot/system drive 480gb

 

WD 2tb mass storage HD

 

GPU/problem child!  

Gigabyte Aorus RX 580 8gb  

(this card was used and had a bad fan when I got it. Replaced bad fan and found even with both fans it was a crazy hit blast furnace so I installed a Arctic Accelero Xtreme  IV which has kept gaming temps at or below 70c max but I have seen the temps run off stress testing up to the 90c before, and it starts to get screen glitches when temps start to hit 80-85c at which I either crashes or I stop the stress test. 

 

Case is a DIYPC TG8-BW

 

so with all that info on the table I have not OC the GPU cause it won’t run a stress test(kombustor) with the factory 1365 clock speed and 2150 Memory speed. The test runs for about 2:30-4:00 minute before it starts to screen flicker and glitch or just crash all together. When I run a stress test in full screen and clock speed only hits about 1180 unless I increase the power limit in afterburner to around +24 then it will hit the 1365 but the temps race up fast and the crash happens sooner. I also get that dame wattman default reset message after I reboot. 

 

Now when I game (mostly FPS like apex,PUBG, Z1 and CSGO) the temps stay at or below 70c and the crash is hit or miss. Z1 seems to be the hardest on GPU temps and I run all games at high settings on a 34” LG ultra wide monitor. 

  

I know there is a lot of info here for a crashing question but I wanted to get all the info I have out so there isn’t a hundred back comment questions for more info and guessing. Thanks for the time and any suggestions or  advice would be greatly appreciate 

 

 

 

[ki8aras]

1 minute ago, Bowkiller84 said:

Hello fellow tech juckies. I have recently rebuilt my rig (3rd gen in 12 years). And I’m having some trouble with my GPU crashing. Hoping for some help. I have tried a handful of ideals already but any help would be appreciated. 

 

So to start here is my rig specs 

 

AMD Ryzen 5 1600 with a stable OC at 3.95ghz (has stress test with aida64 for over 9hours on 3 different Occasions) cooled by a deepcool captain 240 with max temps of 72c and normal load temps in the low 50c

 

g-skill trident Z ddr4 ram 3600mhz OC to 3333 stable as well with the same stress test and times as the CPU

 

MSI B 450 tomahawk motherboard

 

Corsair rm850i PSU (under full load using just under 400w give or take) 

 

kington SSD boot/system drive 480gb

 

WD 2tb mass storage HD

 

GPU/problem child!  

Gigabyte Aorus RX 580 8gb  

(this card was used and had a bad fan when I got it. Replaced bad fan and found even with both fans it was a crazy hit blast furnace so I installed a Arctic Accelero Xtreme  IV which has kept gaming temps at or below 70c max but I have seen the temps run off stress testing up to the 90c before, and it starts to get screen glitches when temps start to hit 80-85c at which I either crashes or I stop the stress test. 

 

Case is a DIYPC TG8-BW

 

so with all that info on the table I have not OC the GPU cause it won’t run a stress test(kombustor) with the factory 1365 clock speed and 2150 Memory speed. The test runs for about 2:30-4:00 minute before it starts to screen flicker and glitch or just crash all together. When I run a stress test in full screen and clock speed only hits about 1180 unless I increase the power limit in afterburner to around +24 then it will hit the 1365 but the temps race up fast and the crash happens sooner. I also get that dame wattman default reset message after I reboot. 

 

Now when I game (mostly FPS like apex,PUBG, Z1 and CSGO) the temps stay at or below 70c and the crash is hit or miss. Z1 seems to be the hardest on GPU temps and I run all games at high settings on a 34” LG ultra wide monitor. 

  

I know there is a lot of info here for a crashing question but I wanted to get all the info I have out so there isn’t a hundred back comment questions for more info and guessing. Thanks for the time and any suggestions or  advice would be greatly appreciate 

 

 

 

have you tried reinstalling the drivers?

 

[Bowkiller84]

6 minutes ago, ki8aras said:

have you tried reinstalling the drivers?

 

I did last night. And still have same problems. 

[bejamartins]

37 minutes ago, Bowkiller84 said:

Hello fellow tech juckies. I have recently rebuilt my rig (3rd gen in 12 years). And I’m having some trouble with my GPU crashing. Hoping for some help. I have tried a handful of ideals already but any help would be appreciated. 

 

So to start here is my rig specs 

 

AMD Ryzen 5 1600 with a stable OC at 3.95ghz (has stress test with aida64 for over 9hours on 3 different Occasions) cooled by a deepcool captain 240 with max temps of 72c and normal load temps in the low 50c

 

g-skill trident Z ddr4 ram 3600mhz OC to 3333 stable as well with the same stress test and times as the CPU

 

MSI B 450 tomahawk motherboard

 

Corsair rm850i PSU (under full load using just under 400w give or take) 

 

kington SSD boot/system drive 480gb

 

WD 2tb mass storage HD

 

GPU/problem child!  

Gigabyte Aorus RX 580 8gb  

(this card was used and had a bad fan when I got it. Replaced bad fan and found even with both fans it was a crazy hit blast furnace so I installed a Arctic Accelero Xtreme  IV which has kept gaming temps at or below 70c max but I have seen the temps run off stress testing up to the 90c before, and it starts to get screen glitches when temps start to hit 80-85c at which I either crashes or I stop the stress test. 

 

Case is a DIYPC TG8-BW

 

so with all that info on the table I have not OC the GPU cause it won’t run a stress test(kombustor) with the factory 1365 clock speed and 2150 Memory speed. The test runs for about 2:30-4:00 minute before it starts to screen flicker and glitch or just crash all together. When I run a stress test in full screen and clock speed only hits about 1180 unless I increase the power limit in afterburner to around +24 then it will hit the 1365 but the temps race up fast and the crash happens sooner. I also get that dame wattman default reset message after I reboot. 

 

Now when I game (mostly FPS like apex,PUBG, Z1 and CSGO) the temps stay at or below 70c and the crash is hit or miss. Z1 seems to be the hardest on GPU temps and I run all games at high settings on a 34” LG ultra wide monitor. 

  

I know there is a lot of info here for a crashing question but I wanted to get all the info I have out so there isn’t a hundred back comment questions for more info and guessing. Thanks for the time and any suggestions or  advice would be greatly appreciate 

 

 

 

I have the same GPU, although from Saphire and drivers are a big issue. They go up and down in a rollercoaster. I don't even remember the last fully stable driver I had. Since 19.4.1 for me it's been ok-ish, although some games like AC Odyssey crash with them. 19.4.0 crashed the entire PC, due to messed up power configuration.

 

It's the first AMD card I bought in a long time, and honestly with this quality of support, I think I'll go up with the extra cash and go team green next time.

[Bowkiller84]

1 hour ago, bejamartins said:

I have the same GPU, although from Saphire and drivers are a big issue. They go up and down in a rollercoaster. I don't even remember the last fully stable driver I had. Since 19.4.1 for me it's been ok-ish, although some games like AC Odyssey crash with them. 19.4.0 crashed the entire PC, due to messed up power configuration.

 

It's the first AMD card I bought in a long time, and honestly with this quality of support, I think I'll go up with the extra cash and go team green next time.

I am running 19.4.1 and having the same problem as the .0 but I may go back and try a older driver this weekend for the hell of it and see if that helps.

[bejamartins]

1 hour ago, Bowkiller84 said:

I am running 19.4.1 and having the same problem as the .0 but I may go back and try a older driver this weekend for the hell of it and see if that helps.

If you find a sweet spot, please tell me as well. I won't be near that PC for a few days. If meanwhile I have the opportunity to test it first I'll share as well. Someone (we?) should create a "tested" table so people have a reference, else it's always a nightmare.

[Bowkiller84]

1 hour ago, bejamartins said:

If you find a sweet spot, please tell me as well. I won't be near that PC for a few days. If meanwhile I have the opportunity to test it first I'll share as well. Someone (we?) should create a "tested" table so people have a reference, else it's always a nightmare.

Just looking over the driver details the past few have really only been Vega support from what I can tell

[Bowkiller84]

Well I used DDU and removed all old drivers and reinstalled the latest one and thought for a few minutes I fixed the problem. But at 6:30 minutes of running stress test I had a soft crash. Screen flashed white then back to normal with blue and purple dots all over then back to normal image but needed to restart to get any program to run right. Temp was only at 81c so it didn’t over heat so I don’t know what to think other then a maybe shopping for a new GPU

[bejamartins]

20 hours ago, Bowkiller84 said:

Well I used DDU and removed all old drivers and reinstalled the latest one and thought for a few minutes I fixed the problem. But at 6:30 minutes of running stress test I had a soft crash. Screen flashed white then back to normal with blue and purple dots all over then back to normal image but needed to restart to get any program to run right. Temp was only at 81c so it didn’t over heat so I don’t know what to think other then a maybe shopping for a new GPU

My love with my RX was short lived as well.

[ThePD]

Try an undervolt. Maybe there is some weird power issue with the card. Why did you buy the card damaged?

[Cosmicjellyfish]

i run with the xfx 580 8g gt oc+ black edition. i keep temps to 70c and when overclocking i hit 81c. my case is trash and i blew the front fan wiring wrong. i'm using the 19.4.1 update im going to try doing crossfire because all my games can utilize and ive been really happy with this purchase. 80c isnt bad for the card id just worry about hitting 84c+ for a long duration.

[Bowkiller84]

16 hours ago, ThePD said:

Try an undervolt. Maybe there is some weird power issue with the card. Why did you buy the card damaged?

That was my next move. I set the core clock to the RX reference clock of 1340 and put the power limit to -10 and this are much more stable now. And I didn’t buy it knowing the fan was bad. I have to do more testing on it and I have a couple more ideals to try but I’m thinking of making a jump to a brand new in the box unopened Sapphire nitro+ RX 590 in a couple months if I can’t get anywhere with this 580

[Peskanova]

Dont undervolt, try to set power limit at +50% (from radeon wattman you can lower power limit from -50 to +50%) with that you tell to the gpu that pick a stable frequency and up your power when it needs, (making a stable frequency without spikes) 

had the same issue with a rx 480 , and now 0 spikes since then
 

 

[Bowkiller84]

Some thing that has come to mind is that maybe and most likely this card was used for mining and the last owner could have done a mining bios flash on the card? Anyone think that maybe choking the power and not letting the card run full bore on games and stress test???

[Bowkiller84]

Update!!!  So this morning I flashed my GPU BIOS to the latest one released that I could find on tech power up for the RX 580 8gb and it seems to have help if not fixed the problems I was having. I have ran a kombustor stress test at default setting for over 20 minutes with NO artifacting, glitching or crashes.  Now I will start to play with overclocking but if it can’t I’m still happy the card is able to run stable at stock now. 

[7Hertz]

Running sapphire RX 570 and not a single issue with any drivers, this has to be something else.

The card might just be a dud.

[Bowkiller84]

After spending all day benchmarking and stress testing my system both CPU and GPU I can say with confidence that the GPU problem was a bad BIOS. Now I have the GPU OC 800mhz and stable with good air cooling temps. Thanks everyone that pitched in ideals to help get this card fixed!

[bejamartins]

On 4/20/2019 at 12:58 PM, Bowkiller84 said:

Update!!!  So this morning I flashed my GPU BIOS to the latest one released that I could find on tech power up for the RX 580 8gb and it seems to have help if not fixed the problems I was having. I have ran a kombustor stress test at default setting for over 20 minutes with NO artifacting, glitching or crashes.  Now I will start to play with overclocking but if it can’t I’m still happy the card is able to run stable at stock now. 

 

Sorry for the late reply. Where can you get updated BIOS?

[LOST TALE]

I've had mine for 9 months and I have to lower memory oc by 5mhz every 2 months to return to stability. OCing isn't what it used to be.

 

Also could barely oc beyond stock 1365mhz. bc it would exceed the 225w power limit

[c00face]

5 hours ago, bejamartins said:

 

Sorry for the late reply. Where can you get updated BIOS?

What this guy said.

 

Sapphire does not release important vBIOS update on their cards and it's causing major crashes. I don't know if this is just to save company-face and pretend like nothing is going on. But you're not the only one suffering from these BSOD, hang, crashes on these cards. There are so many threads related to this all over tech forums and I'm surprise nobody has put it together yet or even call out AMD.

 

Just like above user, I had to undervolt my card just to get it to stop repeatedly crashing after running a stress test. I'm still waiting on Sapphire tech support to give me my requested vBIOS update on the GPU.

[bejamartins]

11 hours ago, c00face said:

What this guy said.

 

Sapphire does not release important vBIOS update on their cards and it's causing major crashes. I don't know if this is just to save company-face and pretend like nothing is going on. But you're not the only one suffering from these BSOD, hang, crashes on these cards. There are so many threads related to this all over tech forums and I'm surprise nobody has put it together yet or even call out AMD.

 

Just like above user, I had to undervolt my card just to get it to stop repeatedly crashing after running a stress test. I'm still waiting on Sapphire tech support to give me my requested vBIOS update on the GPU.

 

Damn...

 

I guess I can invoke warranty and demand my money back.

[Dinner_bone]

On 4/20/2019 at 6:07 PM, Bowkiller84 said:

After spending all day benchmarking and stress testing my system both CPU and GPU I can say with confidence that the GPU problem was a bad BIOS. Now I have the GPU OC 800mhz and stable with good air cooling temps. Thanks everyone that pitched in ideals to help get this card fixed!

How do you fix the bios?

[Goildzy]

Run ur ryzen at 3.7 and ir below 1.25 voltsge thank me later

[Bowkiller84]

Sorry for not replying sooner. Haven’t been on in awhile. So to answer the where and how did I update the bios. I didn’t really update I flashed a new bios or flashed back to the old bios. The card I had was a used card and used for mining and there was a mining bios flashed onto the card which made it crunch the help out of numbers but crap on gaming graphics. So I used a site called Techpowerup to find the right bios and then used ATIFlash to flash the new bios onto the card. Just know that if anything goes wrong when flashing the bios then the card will Brick and be a paperweight. Just a warning. Good luck

[Bowkiller84]

And Goildzy my OC on the system is 3.95ghz, memory is at 3333mhz and I know I’m under voltage but can’t think of what it’s at off the top of my head. But the system is very stable at those sittings.